R Notebook

Read Data

#install.packages("sqldf")    #https://cran.r-project.org/web/packages/sqldf/sqldf.pdf   Permite realizar búsquedas sobre dataframes con sintaxis sql
library(sqldf)
#install.packages("fmsb")  #Several utility functions for the book entitled "Practices of Medical and Health Data Analysis using R" (Pearson Education Japan, 2007) with Japanese demographic data and some demographic analysis related functions.
library(fmsb)
library(plotly)  #https://plot.ly/r/
video_games<-read.csv("recursos/vgsales.csv")
typeof(video_games)

[1] "list"

head(video_games)

video_games$Year <- ordered(video_games$Year)
video_games <- video_games[video_games$Year<2017,]
video_games

Explore Sales by Year and Platform

ByYear<-sqldf("SELECT Platform, Year, sum(NA_Sales) as AME, sum(EU_Sales) as EU,sum(JP_Sales) as JP, sum(Other_Sales) as Other, sum(Global_Sales) as Global from video_games group by Platform, Year order by Year,Platform")

Ventas anuales. Todas las Plataformas

All<- sqldf("SELECT Year, sum(NA_Sales) as AME, sum(EU_Sales) as EU,sum(JP_Sales) as JP, sum(Other_Sales) as Other, sum(Global_Sales) as Global from df group by Year order by Year")
#https://plot.ly/r/line-charts/
plot_ly(All,type = 'scatter', mode = 'lines') %>%
  add_trace(y=All$Global,x=All$Year,name='Global Sale',mode = 'lines')%>%
  add_trace(y=All$AME,x=All$Year,mode = 'lines',name='American Sale')%>%
  add_trace(y=All$EU,x=All$Year,mode = 'lines',name='European Sale')%>%
  add_trace(y=All$JP,x=All$Year,name='Japen Sales',mode = 'lines')%>%
  add_trace(y=All$Other,x=All$Year,name='Other Sale',mode = 'lines')%>%
  layout(title = "Ventas de Videjuegos Anuales por Región",
         xaxis = list(title = "Años"),
         yaxis = list (title = "Ventas en Millones"))

Lo mismo pero solo para un par de tipos de consola: PS2, PS3 y PS4.

PS2<-subset(ByYear,Platform=="PS2",select = Platform:Global)
PS3<-subset(ByYear,Platform=="PS3",select = Platform:Global)
PS4<-subset(ByYear,Platform=="PS4",select = Platform:Global)
plot_ly(ByYear,type = 'scatter', mode = 'lines') %>%
  add_trace(y=PS2$Global,x=PS2$Year,name='Global Sale PS2',mode = 'lines+markers') %>%
  add_trace(y=PS3$Global,x=PS3$Year,name='Global Sale PS3',mode = 'lines+markers') %>%
  add_trace(y=PS4$Global,x=PS4$Year,name='Global Sale PS4',mode = 'lines+markers') %>%
  layout(title = "Evolución Ventas PlayStation",
         xaxis = list(title = "Años"),
         yaxis = list (title = "Ventas en Millones"))

Detalle PS2 por año y región

PS2<-ByYear[ByYear$Platform=='PS2',]
plot_ly(PS2, type = 'bar', name = 'Ventas PS2 por Región') %>%
  add_trace(y=PS2$AME,x=PS2$Year,name='America',mode = 'bar') %>%
  add_trace(y=PS2$EU,x=PS2$Year,name='Europa',mode = 'bar') %>%
  add_trace(y=PS2$JP,x=PS2$Year,name='Japan',mode = 'bar') %>%
  add_trace(y=PS2$Other,x=PS2$Year,name='Other',mode = 'bar') %>%
  add_trace(y=PS2$Other,x=PS2$Year,name='Global',mode = 'bar')%>%
  layout(title = "Ventas PS2 por año y Región",
         scene = list(
           xaxis = list(title = "Año"), 
           yaxis = list(title = "Ventas")))

PS3<-ByYear[ByYear$Platform=='PS3',]
plot_ly(PS3, type = 'bar', name = 'Ventas PS3 por Región') %>%
  add_trace(y=PS3$AME,x=PS3$Year,name='America',mode = 'bar') %>%
  add_trace(y=PS3$EU,x=PS3$Year,name='Europa',mode = 'bar') %>%
  add_trace(y=PS3$JP,x=PS3$Year,name='Japan',mode = 'bar') %>%
  add_trace(y=PS3$Other,x=PS3$Year,name='Other',mode = 'bar') %>%
  add_trace(y=PS3$Other,x=PS3$Year,name='Global',mode = 'bar')%>%
  layout(title = "Ventas PS3 por año y Región",
         scene = list(
           xaxis = list(title = "Año"), 
           yaxis = list(title = "Ventas")))

PS4<-ByYear[ByYear$Platform=='PS4',]
plot_ly(PS4, type = 'bar', name = 'Ventas PS4 por Región') %>%
  add_trace(y=PS4$AME,x=PS4$Year,name='America',mode = 'bar') %>%
  add_trace(y=PS4$EU,x=PS4$Year,name='Europa',mode = 'bar') %>%
  add_trace(y=PS4$JP,x=PS4$Year,name='Japan',mode = 'bar') %>%
  add_trace(y=PS4$Other,x=PS4$Year,name='Other',mode = 'bar') %>%
  add_trace(y=PS4$Other,x=PS4$Year,name='Global',mode = 'bar')%>%
  layout(title = "Ventas PS4 por año y Región",
         scene = list(
           xaxis = list(title = "Año"), 
           yaxis = list(title = "Ventas")))

Ventas globales por plataforma (principales plataformas)

Platform_data<- sqldf("SELECT Platform, sum(NA_Sales) as AME, sum(EU_Sales) as EU,sum(JP_Sales) as JP, sum(Other_Sales) as Other, sum(Global_Sales) as Global from video_games where platform in('DS','PS','PS2','PS3','Wii','X360','PSP','PS4') group by Platform order by Platform")
Platform_data

plot_ly(Platform_data, labels = ~Platform, values = ~Global, type = 'pie') %>%
  layout(title = 'Ventas totales por consola',
         xaxis = list(showgrid = TRUE, zeroline = TRUE, showticklabels = TRUE),
         yaxis = list(showgrid = TRUE, zeroline = TRUE, showticklabels = TRUE))

plot_ly(Platform_data, labels = ~Platform, values = ~EU, type = 'pie') %>%
  layout(title = 'Ventas totales por consola',
         xaxis = list(showgrid = TRUE, zeroline = TRUE, showticklabels = TRUE),
         yaxis = list(showgrid = TRUE, zeroline = TRUE, showticklabels = TRUE))

Ventas globales por genero

Genre_data<- sqldf("SELECT Genre, sum(NA_Sales) as AME, sum(EU_Sales) as EU,sum(JP_Sales) as JP, sum(Other_Sales) as Other, sum(Global_Sales) as Global from video_games  group by Genre order by Platform")
Genre_data

plot_ly(Genre_data, x = ~Genre, y = ~Global, type = 'bar', text = text,
        marker = list(color = 'rgb(158,202,225)',
                      line = list(color = 'rgb(8,48,107)',
                                  width = 0.5))) %>%
  layout(title = "Ventas Globales por Genero.",
         xaxis = list(title = "",size = 40,color = "#7f7f44"),
         yaxis = list(title = ""))

plot_ly(Genre_data, x = ~Genre, y = ~AME, type = 'bar', text = text,
        marker = list(color = 'rgb(158,202,225)',
                      line = list(color = 'rgb(8,48,107)',
                                  width = 0.5))) %>%
  layout(title = "Ventas Globales por Genero.",
         xaxis = list(title = "",size = 40,color = "#7f7f44"),
         yaxis = list(title = ""))

plot_ly(Genre_data, x = ~Genre, y = ~EU, type = 'bar', text = text,
        marker = list(color = 'rgb(158,202,225)',
                      line = list(color = 'rgb(8,48,107)',
                                  width = 0.5))) %>%
  layout(title = "Ventas Globales por Genero.",
         xaxis = list(title = "",size = 40,color = "#7f7f44"),
         yaxis = list(title = ""))

plot_ly(Genre_data, x = ~Genre, y = ~JP, type = 'bar', text = text,
        marker = list(color = 'rgb(158,202,225)',
                      line = list(color = 'rgb(8,48,107)',
                                  width = 0.5))) %>%
  layout(title = "Ventas Globales por Genero.",
         xaxis = list(title = "",size = 40,color = "#7f7f44"),
         yaxis = list(title = ""))

Por plataforma y Genero

Genre_plat_data<- sqldf("SELECT Platform, Genre, sum(NA_Sales) as AME, sum(EU_Sales) as EU,sum(JP_Sales) as JP, sum(Other_Sales) as Other, sum(Global_Sales) as Global from video_games group by Platform, Genre order by Platform")
gpd_ps4<-Genre_plat_data[Genre_plat_data$Platform=='PS4',]
gpd_ps4

plot_ly(gpd_ps4, labels = ~Genre, values = ~Global) %>%
  add_pie(hole = 0.6) %>%
  layout(title = "Ventas Ps4 por Género",  showlegend = F,
         xaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE),
         yaxis = list(showgrid = FALSE, zeroline = FALSE, showticklabels = FALSE))